Crystallisation, one of the oldest unit operations, is of major economic importance in the hydrometallurgical industry for the separation and/or production of inorganic metal compounds in the form of good quality crystals, both in terms of purity and size. However, industrial crystallisation techniques usually have high energy requirements (e.g. evaporative crystallisation) and the residual solutions are often extremely corrosive, resulting in elevated operating and maintenance costs. Additionally, the majority of industrial crystallisation methods do not render to the appropriate supersaturation control, therefore the produced salts are agglomerates of fine crystallites with heavy solution entrainment and contamination. / In this work, the Solvent Displacement Crystallisation (SDC) technique is investigated as an attractive alternative to the conventional crystallisation methods. SDC involves the addition of low-boiling point, water-miscible organic solvents (MOS) to aqueous solutions to cause salt precipitation based on the salting out effect. The crystals are separated by filtration whereas the solvent is subsequently recovered for reuse by low-temperature distillation. The present work was initiated with the main objective to develop a solid scientific understanding of the SDC process and propose specific applications to hydrometallurgical systems of practical interest. / Criteria for the selection of organic solvents with suitable physical and chemical properties have been established and various compounds screened to determine their amenability to SDC; 2-propanol was selected as an effective salting out agent to cause precipitation of most metallic sulphates of practical interest from acidic solutions and opted for use in further studies. Differences in crystallisation behaviour among the various metal sulphates were attributed to differences in hydration energy (the energy required for a hydrated ion to be separated from its bound water). / None of the tested metal chlorides could be successfully separated from HCl-H2O system with 2-propanol. This was explained in terms of enhanced metal chlorides solubilities in non-aqueous solvents relative to water by formation of chloro-complexes of larger stability constants. The preferential formation of chloro-complexes in mixed aqueous and organic solvents is the result of the almost linear drastic increase in the activity of Cl - with mole fraction organic. / Successful examples of using the SDC method in conjunction with an industrial process involve precipitation of NiSO4·6H2O from copper electrorefining spent electrolytes, residual sulphate removal as gypsum (CaSO4·2H2O) in chloride leaching processes and ZnO separation/NaOH regeneration in the system Na2ZnO2, - NaOH - H2O. By maintaining a low supersaturation (i.e. controlled addition of the solvent to the electrolyte) and heterogeneous crystallisation conditions (use of seed/product recycling), crystal growth is favoured while impurity uptake/contamination is minimised.

Several mineral flotation plants utilize process water with high soluble salt content. The processing of soluble minerals, the use of recycle streams, sea or well water give rise to this situation. It is unclear as to the overall effect on flotation response though there does seem to be evidence of bubble size reduction, increased froth stability and reduced reagent dosage. The present work aims to quantify the effects of salts commonly present in process water on gas dispersion (bubble size and gas holdup) and froth properties (solid and liquid overflow rates) and to compare to a typical frother, MIBC (methyl isobutyl carbinol). Three sets of experiments were run: 2-phase, 3-phase using talc and 3-phase using ore. All tests were run in a laboratory column. The salts examined in the 2-phase tests were NaCl, CaCl₂, Na₂SO₄, Na₂S₂O₃ and Al₂(SO₄)₃. Tests were conducted to determine the effects of salt type and concentration on gas holdup, bubble size distribution and foaming. The salts containing multivalent ions had a greater impact than did the monovalent ions, in keeping with the literature. A relationship between gas holdup and ionic strength was established. The effects of NaCl and MIBC on frothing and on solid and liquid overflow rates in 3-phase tests using 1%w/w talc were investigated. The results showed increased frothing and overflow rates upon the addition of NaCl, comparable to adding MIBC. Three-phase tests using ore (Falconbridge's Brunswick Mine) explored the effect of NaCl concentration. The results show reduction in bubble size and increased gas holdup in salt solutions comparable to those in the 2-phase tests. For each case, the concentration of MIBC giving equivalent gas holdup and frothing as the salt systems was determined. Proposed mechanisms for coalescence inhibition and froth stability in salt solutions are briefly reviewed in light of the current findings.

Zirconium alloys pickup hydrogen, during reactor operation. When the hydrogen content exceeds the solubility limit, the hydrides are formed. These hydrides have deleterious influence on the mechanical properties of the material. Therefore, it is important to improve our understanding of hydride morphology and behavior at different conditions approximating nuclear reactor environments. Initial analysis was done on the hydrided, cold-worked and stress-relieved (CWSR) Zircaloy-4 sheet-samples. The hydrides have been identified as δ-ZrH1.5 phase. They were located within the grains and along the grain boundaries, but the grain boundary hydrides appeared to be dominant. The preferential orientation relationship of (0001)α-Zr//(111)δ-ZrH1.5 was observed between hydrides and the matrix at all locations. The changes in hydrogen concentration had no effect on the crystallographic orientation relationship of the hydrides. Subsequently, the hydride behavior under hoop stress has been studied in detail. A range of hoop stresses were applied to the Zircaloy-4 cladding tube in a high pressure hydrogen gas furnace at a temperature of 350ºC and a pressure of 20 MPa. Optical micrographs of the samples showed an increase in the number of hydrides oriented along radial direction at higher hoop stresses. Apart from radial hydrides, the hydrides which were inclined at an angle to circumferential and radial directions were also observed. Interestingly, the hydrides which were oriented perpendicular to the circumferential direction followed a {10 1}α-Zr//{111}δ-ZrH1.5. relationship with the matrix grains. Habit planes of {10 1}, {10 7}, {10 3}, {10 0} were observed for the hydrides which were inclined at an angle to the circumferential direction. This study helps in understanding the complex behavior of radial hydrides. An attempt has been made to understand the superior resistance to corrosion and substantially lower hydrogen pickup of the Zr-1Nb alloy when compared with Zircaloy-4. Thus, Zircaloy-4, Zr-2.5Nb and Zr-1Nb (E110) alloys were hydrided in a high pressure hydrogen gas furnace. The hydrides in Zr-1Nb (E110) were short and thick, without any preferential orientation with respect to the sample reference direction. Heavy {10 2} twins were observed in the Zr-1Nb alloys, which were not seen in the Zircaloy-4 and Zr-2.5Nb alloys. Almost 50 % of hydrides in the Zr-1Nb (E110) alloy showed a new crystallographic relation of (0001)α-Zr//{100}δ-ZrH1.5. This new orientation relationship in the Zr-1Nb alloy is attributed to the presence of twin boundaries. In addition, the microstructural changes at the weld zone and its effects on the hydride behavior have been studied. The weld zone shows predominantly basket-weave structures, with very few colonies of parallel plate structures. Variant selection for α-phase formation inside prior β-grains was identified at the weld center. As we move from the weld center to the as-received zone, the variant selection is found to be less probable. At the weld zone, a complete random orientation of hydrides was observed. Moreover, the δ-hydride platelets at the weld zone were always found to be growing perpendicular to the α-colonies, which had a misorientation of 60º, and followed a (0001)α-Zr//{111}δ-ZrH1.5 orientation relationship with the zirconium matrix. A proposed description of the complex distribution of hydrides and alloy microstructures, at the weld and heat affected zones, will contribute to a better understanding of the mechanisms of fuel cladding failure in various types of nuclear reactors. / Les alliages de zirconium ramassent de l'hydrogène, au cours de l'opération de réacteur. Lorsque la teneur d'hydrogène dépasse la limite de solubilité, les hydrures sont formés. Ces hydrures ont un effet néfaste sur les propriétés mécaniques du matériel. Par conséquent, il est important d'améliorer notre compréhension de la morphologie des hydrures et leur comportement sur les conditions différentes rapprochant des environnements de réacteur nucléaire. L'analyse initiale a été faite sur les hydrided hydrures, travaillés à froid et soulagés de contraintes du Zircaloy-4 (CWSR) feuille-échantillons. Les hydrures ont été identifiés comme phase δ-ZrH1.5. Ils sont trouvés dans les grains et le long des frontières de grains, mais les hydrures de limite de grain semblaient être dominants. La relation d'orientation préférentielle de (0001)α-Zr// (111)δ-ZrH1.5 a été observée entre hydrures et la matrice à tous les endroits. Les variations de concentration de l'hydrogène n'ont eu aucun effet sur la relation de l'orientation cristallographique des hydrures. Par la suite, le comportement d'hydrure sous hoop stress a été étudié en détail. Une plage de hoop stress a été appliquée au tube de revêtement du Zircaloy-4 dans un four à gaz à haute pression d'hydrogène à une température de 350ºC et sous une pression de 20 MPa. La microscopie optique des échantillons ont montré une augmentation du nombre des hydrures orientés le long de la direction radiale à des données de hoop stress plus élevées. Hormis les hydrures radiaux, les hydrures qui étaient enclins à un angle de circonférence et radiales des instructions ont aussi été observés. Fait intéressant, les hydrures qui étaient orientés perpendiculairement à la direction de circonférence suivent la relation de {10 1}α-Zr//{111}δ-ZrH1.5 avec les grains de la matrice. Les plans d'habitude de {10 1}, {10 7}, {10 3}, {10 0} ont été observés pour l'hydrure qui étaient enclins à un angle avec la direction de la circonférence. Cette étude permet de comprendre le comportement complexe des hydrures radiaux. Une tentative à comprendre la résistance à la corrosion et à baisser substantiellement la ramasse de l'hydrogène de l'alliage Zr-1Nb comparativement à Zircaloy-4. Ainsi, Zircaloy-4, Zr-2.5Nb et alliages Zr-1Nb (E110) ont été hydrudés dans un four à gaz à haute pression de l'hydrogène. Les hydrures dans le Zr-1Nb (E110) étaient courts et épais, avec aucune orientation préférentielle par rapport à la direction des échantillons de référence. Des jumeaux lourds {10 2} ont été observés dans les alliages de Zr-1Nb, qui n'étaient pas considérés dans la Zircaloy-4 et Zr-alliages 2.5Nb. Près de 50 % des hydrures dans l'alliage Zr-1Nb (E110) ont montrés une nouvelle relation cristallographique de (0001)α-Zr//{100}δ-ZrH1.5 Cette nouvelle relation d'orientation dans l'alliage Zr-1Nb est attribuée à la présence des limites des jumeaux. En outre, les changements des microstructures à la zone de soudure et ses effets sur le comportement d'hydrure ont été étudiés. La zone de soudure montre surtout des structures nattées, avec très peu de colonies ayant des structures de plaques parallèles. La sélection variante pour la formation de α-phase à l'intérieur de grains de β préalables a été identifiée au centre de la soudure. Alors que nous déplaçons du centre de soudure à la zone reçu, la sélection de la variante est jugée moins probable. Dans la zone de soudure, une orientation aléatoire complète des hydrures a été observée. En outre, les plaquettes δ-hydrure dans la zone de soudure ont toujours été croissante perpendiculaire aux α-colonies, qui avaient une différence angulaire de 60º et qui suivent une (0001)α-Zr//{111}δ-ZrH1.5 relation d'orientation avec la matrice de zirconium.

A novel method has been developed to prepare magnetic carriers for metal ion removal from dilute solutions. In this work, magnetic carriers were prepared by coating an organic surfactant on magnetic particles. The coating was deposited by molecular self-assembly. Two coatings were tested, using two similar bolaamphiphiles* of different chain lengths, HS-(CH2)11-COOH and HS-(CH 2)15-COOH. The magnetic particles were nanosized maghemite (gamma-Fe2O3). / The surfactants were synthesized and characterized. Magnetic carriers were prepared, characterized, and tested for metal ion loading. Molecular orientation, density, and stability of the surfactant coatings were characterized by diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS), x-ray photoelectron spectroscopy (XPS), wetting tests, and leaching tests. The coatings, resulting from the chemisorption of the surfactants on maghemite through carboxylate bonding, were found to be dense, well-packed, and resistant to acid and base attack. / The potential use of coated particles for metal ion removal was assessed with Cu and Ag solutions. The metal ion uptake by the carriers was characterized by atomic adsorption and MRS. The influence of pH and metal concentration on adsorption was studied. Metal ion uptake was found to be dependent on pH, and was not significantly different for both types of coated carriers and for bare maghemite. The maximum loading capacity was low, at approximately 6 mg Cu/g particles for Cu. The similarity in loading for coated and bare particles needs further clarification. / *Bolaamphiphiles are surfactants with a functional group at each end.

Five slightly different IF 409 ferritic stainless steels were investigated by means of continuous cooling compression (CCC) testing. These compression test were performed at a constant strain rate of $1.8 times10 rm sp{-4}s sp{-1}$ over a 600$ sp circ$C temperature range (i.c. from 1200$ sp circ$C- 600$ sp circ$C) at four different cooling rates (0.25, 0.5. 1.0 and 2.0$ sp circ$C/sec). The reheating temperature and time were varied and the effect of this on precipitation is discussed. / The analysis of the CCC test results, and the optical microscopy observations, indicated that neither solute drag nor phase transformation is responsible for the deviations observed in these flow curves. TEM bright field images were used to verify the presence of precipitates, and to calculate the strengthening influence of these particles on the flow curve. It is shown that, in the niobium bearing materials, the precipitation of carbonitrides is responsible for the increase in flow stress. Conversely, in the titanium bearing materials, the titanium carbonitride precipitates were determined to have no effect on the flow stress. / It was discovered that increasing the solution temperature and time resulted in the occurrence of precipitation at higher temperatures, and that these precipitates were coarsened to a greater extent. Also, when the holding temperature and/or the holding time was decreased, the amount of niobium put into solution decreased, which led to precipitation starting at lower temperatures. / A method was developed for determining the precipitation start and finish temperatures from the CCC results. These temperatures were converted to precipitation start and finish times for each of the four cooling rates. This revealed that the precipitation start and finish times increase with decreasing cooling rate in both the 0.2% Nb and 0.1% Nb + 0.1% Ti stainless steels. Finally, a continuous cooling precipitation diagram was constructed for the two niobium bearing materials using these start and finish times.

The continuous heating to fracture test (CHF) developed at McGill was used to study the effect of decarburization on the incipient melting temperature. In the CHF test, a sample was deformed at a constant true strain rate, while its temperature is increased at a constant rate. This test allows the incipient melting temperature to be determined using a single test, instead of the several required by isothermal tensile testing. An isothermal decarburization step, to create a decarburized layer at the surface of the sample, was added prior to the CHF test. / Analysis of the true stress vs temperature curves obtained by CHF testing allows the incipient melting temperature to be determined. Optical metallography was used to determine the effect of decarburization time on the observed depth of decarburization. The heat transfer characteristics of induction heating were studied, since this heating technique is known to create a significant temperature gradient at the surface of the piece being heated. Since the decarburized layer is in the high temperature zone, and also has a higher melting point due to its lower carbon content, it plays an important role in the melting process. / A phenomenological model was derived to describe the melting process. It allows for both the temperature gradient due to induction heating and the melting point gradient due to decarburization. The hypothesis is advanced that melting takes place at the position, within the sample, where the temperature profile crosses the incipient melting temperature gradient. From this study, it appears that decarburization acts so as to limit the risk of hot shortness on a workpiece being forged at high temperatures. This is because decarburization raises the IMT, and in this way widens the temperature window of optimum workability.

Formation of Pb and Ca precipitates on a silica surface in the presence of sulphate and carbonate under neutral to alkaline condition is studied using zeta potential measurements, Scanning Electron Microscopy (SEM), and X-Ray Photoelectron Spectroscopy (XPS). The effect of carbonate on displacing sulphate is discussed. / In the Pb/sulphate system, precipitates of predominantly Pb-oxide/hydroxide formed on the silica surface. In the Pb/carbonate system, the precipitates were predominantly Pb-carbonate/hydroxy carbonate. Precipitates in both the Pb/SO$ sb4$ and Pb/CO$ sb3$ systems were dispersed by addition of sufficient carbonate, the silica surface eventually becoming free of precipitates. This dispersion was considered to be due to: a high negative surface charge density on both lead carbonate and silica; nucleation being promoted at high concentrations of carbonate, resulting in small precipitates; and increased solubility at high carbonate concentrations. / In the Ca system, calcium sulphate precipitates did not form on the silica particles while calcium carbonate precipitates did. The precipitates were not removed by increasing carbonate concentration. The difference from the Pb system is attributed to a low negative surface charge density on calcium carbonate compared to lead carbonate, leading to less dispersing conditions, and the higher solubility of calcium sulphate compared to lead sulphate.

The onset of a 'late' rotating vortex over an off-centre drain nozzle at 2/3 radius was studied in an 1160-mm diameter tank. It was found that using a sloped bottom ladle could be beneficial in terms of steel yield, provided the exit nozzle is located 'centrically'. / Miner modification of the nozzle (skewed nozzle) to impart a radial component of velocity to the spinning vortex core was found to be effective in making AMEPA system sensitive to early slag entrainment phenomena by diverting the core away from the central vertical axis of the nozzle. / A 0.75 scale water model was constructed to simulate the flow of liquid steel through a ladle shroud in the presence of gas infiltration. It was found that the ladle shroud slag detector could be temporarily 'blinded' by gas bubbles or permanently blinded by a standing submerged gas jet.

Variant selection rules are described that predict the austenite-to-mariensite transformation textures of an Fe-30% Ni alloy. The occurrence of variant selection is explained in detail using the crystal plasticity model of Bishop and Hill and the Kurdjumov-Sachs transformation relationship. A correlation is established between the slip systems of the former and the rotation axes of the latter. The selection criteria are based on a combination of slip activity, that is, active slip systems defined in terms of positive shears, and permissible dislocation reactions. Thus some of the variants selected are associated with slip systems that are active, while the remainder are accounted for by the in-plane reaction of active dislocations to form inactive or unstressed dislocations. The variant selection criteria are tested against two strain paths; plane strain rolling and axisymmetric compression. Extensive studies were conducted on the former strain path and revealed excellent agreement between the measured and predicted transformation textures. A more detailed analysis is still required for the axisymmetric compression strain path, although, the results obtained indicate that the variant selection model does indeed generate a transformation texture that is consistent with the experimental one.

A full-scale physical model of a delta-shaped, four-strand tundish was constructed at McGill University to find optimal operating conditions for the tundish. The objective of this work was to determine how best to increase steel production rates by 14% over current tundish operating conditions, these being 12 tons/min with four 15.0 mm outlet nozzles for a 500 mm depth of liquid steel while maintaining steel quality levels. Two options were suggested: the normal head option uses 16.0 mm outlet nozzles and maintains the 500 mm tundish level, while the high head option uses 14.8 mm outlet nozzles and raises the depth of liquid steel within the tundish to 800 mm. The important effects of flow control devices on the hydrodynamic performance of the tundish were also tested, using two different types of flow modifiers: Impact Pad and Turbo-Stop. For a proper comparison between the two options, three aspects were investigated; vortex formation phenomena during tundish draining between ladle changes, Residence Time Distribution (RTD), and Inclusion Separation Ratios (ISR). Inclusion removal rates were studied experimentally with the aid of the aqueous" Liquid Metal Cleanliness Analyzer (LiMCA) system. Particle Image Velocimetry (PIV) was used to visualize the actual instantaneous, or momentary, flows, thereby providing the data needed for time averaged velocity fields and turbulent kinetic energies. A mathematical model based on METFLO was developed to simulate these tundish operations numerically. The Renormalization Group turbulence model (RNG) as well as the standard high Reynolds number k-s turbulence model (STD) was implemented in order to simulate the turbulent flows within the tundish. The validity of METFLO was confirmed by PIV measurements and the numerical predictions of the RTD, and RRI matched the results of physical modelling.